Research On Optimization And Evaluation Method Of Selective Maintenance Strategy Based On Swarm Intelligence Algorithm And Fuzzy Multi-attribute Decision Making

With the active promotion of "made in China 2025" strategy,the traditional equipment manufacturing industry has gradually transformed and upgraded to a new manufacturing industry with intelligent,lean,large-scale and green development as the core.As an essential part of equipment life cycle,maintenance support plays an increasingly important role.Although the scientific and technological innovation has brought about a significant improvement in the level of equipment manufacturing,the maintenance support problems of equipment system come one after another,especially the contradiction between limited maintenance resources and maintenance support decision-making is increasingly prominent.In practical engineering,the maintenance decision-making usually takes into account the existing resource constraints,and selectively maintains some components of the whole system,that is,selective maintenance.It is the key to study the maintenance decision-making problem under resource constraints and put forward a reasonable maintenance resource optimal allocation scheme,then achieve the preset goal reliably.Therefore,it is of great theoretical and practical significance to carry out research on selective maintenance strategies and evaluation methods under resource constraints.Based on the theory of selective maintenance,swarm intelligence optimization algorithm and multi-attribute decision-making technology,and on the premise of expanding the basic concept and scope of selective maintenance,this paper constructs decision-making models of selective maintenance for some problems found in the field of selective maintenance,and puts forward new algorithms of selective maintenance strategy optimization and quality evaluation.The validity of the model and algorithm is verified by examples.The main research contents and beneficial results of this paper include the following aspects:(1)Research on single objective selective maintenance strategy optimization based on DGSA algorithmThe energy efficiency index is introduced into the selective maintenance decision-making problem,the single objective selective maintenance optimization problem is studied.The Kijima ? imperfect maintenance model is introduced to describe the imperfect maintenance characteristics of multi-state system,the task reliability model of series parallel system is established,and the energy consumption model of selective maintenance process is established around the energy consumption stage of maintenance process.A differential-gravity search algorithm(DGSA)combining differential operators is designed,which is applied to two different scale examples to generate the Pareto schemes efficiently.By comparing DGSA with differential algorithm and gravitation search algorithm,the effectiveness of the proposed algorithm is verified.(2)Research on multi-objective selective maintenance strategy planning based on two-stage methodAiming at the matching problem between system task requirements and team maintenance capability,this paper puts forward the concept of team maintenance capability factor,analyzes the structural characteristics of hybrid parallel system,and establishes a multi-objective selective maintenance optimization model considering team maintenance capability under time constraints,that is,the maximum system reliability and the minimum maintenance cost.A hybrid solution method of MADM and intelligent algorithm is proposed,that is,on the basis of MADM evaluation of team maintenance capability,dynamic multi-objective artificial bee colony algorithm is adopted to calculate Pareto frontier of multi-objective selective maintenance model.The example shows that Pareto solution under different team maintenance capability are generated,and the general rule of matching task requirements and team maintenance capability is explored.Finally,the effectiveness of dynamic multi-objective artificial bee colony algorithm is verified by comparing with NSGA-?.(3)Research on multi-objective cooperative selective maintenance sequence planning based on MOGSA algorithmBased on the maintenance level matching optimization of system components,the cooperative selective maintenance sequence planning under task interference is studied for the product maintenance work with multi-person cooperation.An improved hierarchical maintenance tree is used to express the interference relationship between maintenance tasks,and a multi-objective optimization model aiming at maximizing maintenance revenue and minimizing maintenance time is established.The initialization program of the two-level task sequence coding for parallel maintenance is designed,and the multi-objective gravity search algorithm(MOGSA)for cooperative selective maintenance sequence planning is proposed.The feasibility and effectiveness of MOGSA algorithm in solving the problem of cooperative selective maintenance sequence planning are verified by taking a vehicle's afterburner system as an example.(4)Research on the method of selective maintenance quality evaluation based on fuzzy multi-attribute decision makingAiming at the problem of maintenance quality evaluation with a large number of qualitative indexes and the interaction between them,the fuzzy MADM method is proposed,which combines fuzzy number,artificial bee colony algorithm,?-fuzzy measure,Choquet fuzzy integral and TOPSIS method.Aiming at the selection of maintenance quality scheme,a universal evaluation index system of maintenance scheme is established.In fuzzy MADM method,fuzzy number is used to express fuzzy semantics,and artificial bee colony algorithm is used to identify ?-fuzzy measure.Based on the distance data generated by TOPSIS method,each index value is nonlinear integrated by Choquet fuzzy integral.Taking the selection of equipment maintenance quality scheme as an example,the feasibility and effectiveness of the proposed method are verified by comparing with the weighted grey correlation algorithm and the basic TOPSIS method.